Ellipse Fitting Method Based on Pendulum Genetic Algorithm Particle Swarm Optimization for Well Diameter Measurement by Laser Distance Sensor

The disadvantages of ellipse fitting method based on algebraic method include its large sampling range, unclear physical explanation of the parameters of the equation, uncontrollable accuracy, and susceptibility to outliers. Aiming at these problems, this article presents a method called pendulum algorithm to calculate the shortest distance from a point to a curve that meets a given accuracy requirement. Based on geometric principles, the pendulum algorithm splits the curve and search interval and calculates the shortest distance iteratively. Since the traditional particle swarm optimization (PSO) algorithm may fall into local optimal value, this article incorporates the idea of genetic algorithm (GA), presents the GA PSO (GAPSO) algorithm, and increases the randomness and diversity of particles, which avoids falling into local optimal value. The calculation results of the pendulum algorithm are used as the fitness function of the GAPSO, and the pendulum GAPSO (P-GAPSO) hybrid algorithm is proposed to realize the ellipse fitting to the data points. The P-GAPSO hybrid algorithm was tested through simulated and real experiments, and in the real experiment, it is applied to the simulated well diameter measurement process by laser distance sensor in different environments. Simulation and experiments validate the effectiveness, feasibility, and superiority of the pendulum algorithm and P-GAPSO hybrid algorithm.